Abstract
Using multiple regression in which individual profits were taken as dependent variables, some profit functions were derived from carcass records for 1, 258 head of pigs slaughtered in Kumamoto prefecture in 1991. Economic values on days to markets and carcass traits were derived from partial differentiation of the optimal profit function. The profits of individual pigs were obtained using a bio-economic model of a pig farm. When a linear equation was adopted, the standard partial regression coefficients of profit on back fat thickness in center of carcass (BFc), days to market (DAY) and coefficients of variation on back fat thickness (CVBF) were large, while that dressing percentage (DP) and back and loin length (BLL) were small. The ratio of contribution (R2) was 0.67. Since there was a curvilinear relationship between the profit and traits, a quadratic equation was adopted. This caused R2 to increase to 0.85. Only linear terms were, however, taken for DP and CVBF in the model, because they did not have an optimum level. The economic values (Yen head-1) obtained were: -86 (day-1) for DAY, 128 (%-1) for DP, -57 (cm-1) for BLL, -295 (mm-1) for BFc, and -39 (%-1) for CVBF. Although the economic values derived from approximation of non-linear equations to linear forms were not always optimum in the evaluation of an individual pig, the effects on genetic improvement may be relatively small because of low selection intensity in pig breeding structure. The present approach has a disadvantage in that the partial regression coefficients varied in the adopted equation and may be influenced by correlations among traits. It appears, however, to be useful in those Japanese pig breeding situations where unit prices can not be clearly defined and traits had an optimum level.
